PaddlePaddle / Paddle

PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice (『飞桨』核心框架,深度学习&机器学习高性能单机、分布式训练和跨平台部署)
http://www.paddlepaddle.org/
Apache License 2.0
22.26k stars 5.6k forks source link

什么时候支持Mac M1/M2 GPU? #44830

Closed Lawyer-ray closed 4 months ago

Lawyer-ray commented 2 years ago

需求描述 Feature Description

任务目标(请描述你正在做的项目是什么,如模型、论文、项目是什么?)正在使用Mac M1 Pro作为生产机型,具体使用了PaddleOCR与PaddleNLP,目前Paddle在Mac下仅支持CPU模式,PaddleOCR与PaddleNLP运行效率并不出色; 需求场景(请描述你的项目中为什么需要用此功能)加快解决问题,提高生产效率; ;
功能描述(请简单描述或设计这个功能)

替代实现 Alternatives

No response

paddle-bot[bot] commented 2 years ago

您好,我们已经收到了您的问题,会安排技术人员尽快解答您的问题,请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时,您也可以通过查看官网API文档常见问题历史IssueAI社区来寻求解答。祝您生活愉快~

Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the APIFAQGithub Issue and AI community to get the answer.Have a nice day!

dingjiaweiww commented 2 years ago

感谢你的反馈,对于飞桨支持mac M1 上的GPU,我们内部正在评估调研,请问你需要在mac m1 上运行什么模型呢,使用GPU是想测试一下,还是有特殊项目需要运行呢? @Lawyer-ray

James4Ever0 commented 2 years ago

运行模型可以使用gpu或者neural engine 训练模型使用gpu

可以用coreml createml 进行训练和运行 用metal 3进行训练

目前tensorflow和pytorch都已经支持了metal 3

James4Ever0 commented 2 years ago

search for 'gpu' 'opencl' in codebase to find legacy implementations, which are useful for reimplementation in metal 3. existing implementations in tensorflow-metal and torch (mps) might also be helpful. needless to say the metal 3 documentation.

gudh commented 2 years ago

mac m1 怎么安装兼容?

paddle 不支持这个架构,所以 paddlespeech 也不能支持,抱歉

Wish to use paddlespeech for my new project, but because paddle doesn't support mac M1 platform, as a result, paddlespeech fail to run on mac m1

Wonder if mac m1 will be supported in the near future, any updates?

Anyway, thanks for building this framework!

James4Ever0 commented 2 years ago

now paddle supports m1 CPU but not GPU

install paddle for mac m1 cpu

ioiogoo commented 1 year ago

now paddle supports m1 CPU but not GPU

install paddle for mac m1 cpu

Has any solution for m2 CPU or GPU? I can success installed the package with this command pip install paddlepaddle, but when I import paddle, it's return ERROR: flag 'flagfile' was defined more than once (in files '/Users/paddle/xly/workspace/f2bafd01-b80e-4ac8-972c-1652775b2e51/Paddle/build/third_party/gflags/src/extern_gflags/src/gflags.cc' and '/tmp/gflags-20211021-3963-1mi18ai/gflags-2.2.2/src/gflags.cc')

axuu commented 1 year ago

now paddle supports m1 CPU but not GPU install paddle for mac m1 cpu

Has any solution for m2 CPU or GPU? I can success installed the package with this command pip install paddlepaddle, but when I import paddle, it's return ERROR: flag 'flagfile' was defined more than once (in files '/Users/paddle/xly/workspace/f2bafd01-b80e-4ac8-972c-1652775b2e51/Paddle/build/third_party/gflags/src/extern_gflags/src/gflags.cc' and '/tmp/gflags-20211021-3963-1mi18ai/gflags-2.2.2/src/gflags.cc')

same error.

yaowt05 commented 1 year ago

now paddle supports m1 CPU but not GPU install paddle for mac m1 cpu

Has any solution for m2 CPU or GPU? I can success installed the package with this command pip install paddlepaddle, but when I import paddle, it's return ERROR: flag 'flagfile' was defined more than once (in files '/Users/paddle/xly/workspace/f2bafd01-b80e-4ac8-972c-1652775b2e51/Paddle/build/third_party/gflags/src/extern_gflags/src/gflags.cc' and '/tmp/gflags-20211021-3963-1mi18ai/gflags-2.2.2/src/gflags.cc')

downgrade your opencv-python to 4.6.0.66

ajayghiyad commented 1 year ago

guys I was able to run paddlepaddle CPU version on macbook in docker using this config https://github.com/PaddlePaddle/Paddle/issues/54531#issuecomment-1619601905

paddle-bot[bot] commented 4 months ago

Since you haven\'t replied for more than a year, we have closed this issue/pr. If the problem is not solved or there is a follow-up one, please reopen it at any time and we will continue to follow up. 由于您超过一年未回复,我们将关闭这个issue/pr。 若问题未解决或有后续问题,请随时重新打开,我们会继续跟进。